Bao Jie. All rights reserved

Since Sept 24,2001

means added in Jan 2003

• What is Synergetic Neural Network?
• My Thesis on SNN
• Other works on SNN
• Future work
• Research resource for SNN
• Joke of NN
   

What is Synergetic Neural Network (SNN)?



Hermann Haken

- Includes Two layer: Match layer and Competition layer

- Based on Synergetics and is a kind of self-organization system

FAQ for synergetic neural network
 

---

My Thesis on SNN

• 2000-03-21: The Proposal(Chinese)
• Bao Jie. Research on Some Key Issues of Synergetic Neural Network Master Thesis of Hefei University of Technology, 2001
• Index of the Thesis (Chinese Version)
  • Chapter 1 Intrduction
    • 1．1 Synergetics and Synergetic Information Processing
    • 1．2 Basic principle and importance of Synergetic Neural Network
    • 1．3 Review of current research on Synergetic Neural Network
    • 1．4 Organization of this thesis
    • 1．5 Chapter summary
  • Chapter 2 Principle and Structure Analysis of SNN
    • 2．1 Principle of HAKEN Network
    • 2．2 Principle of Competitive Neural Network
    • 2．3 Look SNN form Competitive Net view
    • 2．4 Implement principles for SNN
    • 2．5 Chapter summary
  • Chapter 3 Matching Subnet
    • 3．1 Meaning and construction of Order Parameter
    • 3．2 Reconstuction of Order Parameter
    • 3．3 Expriment on new method of constructing matching layer
    • 3．4 Orderness measure in SNN
    • 3．5 Chapter summary
  • Chapter 4 Competitive Subnet
    • 4.1 Quick HAKEN network
    • 4.2 Quick HAKEN network : Potential Function Analysis
    • 4.3 Quick HAKEN network : Expriement
    • 4.4 Outlook for hardware implement
    • 4.5 Outlook for improving competitive layer
    • 4.6 Chapter summary
  • Chapter 5 Summary and outlook
  • References
• Code for SNN in my thesis
  • Example for Potential Function
  • 2D Potential Function
  • Prototype pattern and adjoint pattern
  • Order Parameter Evolution Equation
  • Image Recognition
  • Balanced Quick Network
  • Unbalanced Quick Network
  • Residual Vector
  • Adjoint Pattern
  • Projection of input vector
  • Calculate Pseudo Inverse
  • New method for Order Parameter
  • Evolution Function
  • Pseudo-inverse cost
  • Ordeness of SNN
• Thesis Abstract (Chinese , English)
• Slide Show of Defense (in Chinese, 373K, English 292K)
• Defense Notes (in Chinese, 2001-06-20)

This work is supported by following funds:

• 2001.1-2004.1 Research of some key issues on synergetic associative memory, supported by MOE-China, Research assistant; HFUT
• 2000.10-2001.7, Research of some key issues on synergetic associative memory, support by Hefei University of Technology Student Research Foundation, Project Coordinator ;
• 1999.1~2001.12. Research of Associative Memory Based on Mean Field Theory, supported by NSF (National Science Foundation) of China. Research assistant; HFUT
   

---

Other works on SNN

• General
  • 2000-10-25: Study on Key issues and application of Synergetic Associative Memory, by Jie Bao, Yixian Wang
  • 2000-04-28: Synergetic Associative Memory by Jun Gao, Jie Bao
  • 2000-04-12: The Pricinples and the experiments of Haken network, in Chinese
  • 2000-03-06: Review of Synergetic Neural Network
  • 2000-01-12: Introduction to Synergetic Information Processing, in Chinese
  • 2000-01-03: Notes on  Information Compression in Biological Systems
• SNN in Optical-Electronic system
• Invaiance: Hough Transformation
• 2000-12-20: Jun GAO, Jie BAO etal. Optical-Electronic Shape Recognition System Based on Synergetic Associative Memory. SPIE Photonics West 2001, Electronic Imaging 2001 Applications of Artificial Neural Networks In Image Processing VI. Jan 2001. San Jose, USA.
• The Chinese Version
• Beer bottle recognition system
• Matching subnet
• 2000-04-18: Notes on Banzhaf, An Energy Function for Specialization
• 1999-12-23: Notes on HuDongliang, Recognition of objects with skew distortion based on synergetics.
• New Measurement for order parameter
• Entropy of order parameter
• Competition subnet
• 2002-06-06: Potential Function Explain of the Quick Algorithm of Synergetic Neural Network
• Quick Haken network:
• Refer to: 2000-05-13: Key Issues of SNN Research
• Also my thesis Chapter 4
• Robust Competition: [ refer to Roust competitive networks(PDF, 94K), M.Schmutz,W.Banzhaf,  Translated to Chinese by Hong Yan]
• Hierarchical Competition: subdomian competition( form input space); local competition of order parameters(dynamics)
• also Refer to Chen Dingguo's thesis
• Comparsion
• SOM
• ESDN -> refer to Thesis proposal
• BM
• Application
• Content-based retrieval
• Image Recognition：Fingerprint, text

also refer to Hong Yan, Research of Haken Model and Comparison with Other Competitive Learning Network.

---

Future work

• Unfinished Problem (Chinese)
• Within the research of hierarchal neural network
• Within the research of Boundary-based Nerual Network

Research resource for SNN

• Reference List (2002-02-07)
• Collection of Some Original Information about Synergetics (1)-(3), 2000-03-13 to 03-16

Joke of NN

• The Love of SNN (2001-05-08, in Chinese)
• Joke (2002-04-07, in Chinese)